oriGiNAl ArTiClE oscillometric blood pressure measurements: differences between measured and calculated mean arterial pressure H.D. Kiers, J.M. Hofstra*, J.F.M. Wetzels Department of Nephrology (464), Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands, *corresponding author: tel.: +31 (0)24-361 47 61, fax: +31 (0)24-354 00 22, e-mail: [email protected] ABsTract Keywords Mean arterial pressure (MAP) is often used as an index of Blood pressure measurement, mean arterial pressure, overall blood pressure. in recent years, the use of automated oscillometry oscillometric blood pressure measurement devices is increasing. These devices directly measure and display MAP; however, MAP is often calculated from systolic IntroductioN blood pressure (sBP) and diastolic blood pressure (dBP) as displayed by the device. Systemic blood pressure (BP) is one of the most important in this study we have analysed measured and calculated cardiovascular risk factors which is amenable for treatment. MAP, obtained by two different oscillometric BP Thus far most long-term epidemiological studies have used measurement devices in two different patient cohorts. BP values based upon auscultatory measurement with a The first cohort included 242 healthy subjects (male 40.5%, mercury sphygmomanometer. With this technique systolic 50±13 years). BP measurements were performed with a blood pressure (SBP) and diastolic blood pressure (DBP) are Welch Allyn 5300P device. We found a small but significant defined by the appearance and disappearance, respectively, difference between measured MAP and calculated MAP of sounds over the brachial artery during deflation of (MAPm-c: -1.8 mmHg, range -5.7 to 12.9 mmHg, p<0.001). the cuff (Korotkoff sounds I and V). Other indices of BP MAPm-c showed a significant, but weak correlation with can be derived from SBP and DBP. Pulse pressure (PP) dBP and sBP. is calculated by SBP – DBP and mean arterial pressure The second cohort included 134 patients with glomerular (MAP) is calculated by DBP + 1/3 PP. diseases (male 63%, 50±14 years). BP measurements were There is an ongoing debate on which of the above-mentioned performed with a dinamap 487210 device. in this group BP parameters is most important in predicting we also observed a small difference between measured cardiovascular risk and renal outcome.1-4 Some studies MAP and calculated MAP (+1.7 mmHg, range –15.3 to 28.2 suggest that MAP may be more accurate in predicting 1,2 mmHg, p<0.001). MAPm-c correlated with age, all blood cardiovascular prognosis than other BP indices. pressure indices and heart rate. Both in clinical research and clinical practice, the use of An overall analysis showed that age, sBP, dBP, and type of oscillometric BP measurement devices for determining 5 device are all independently related to MAPm-c. BP is increasing. The oscillometric BP measurement There is a significant difference between measured and device measures oscillations from the blood vessel calculated MAP. The difference is small on average; wall during cuff deflation. The pressure at which the however, this MAPm-c can be large in the individual patient. oscillations are maximal is defined as MAP. The device Moreover, there are differences of reported MAP between then calculates the SBP and DBP with an algorithm.6,7 devices. our data suggest that calculated and measured The MAP measured oscillometry is the most reliable MAP cannot be used interchangeably. BP index of the oscillometric BP measurement device.6 Although the measured MAP is reported by most devices, © 2008 Van Zuiden Communications B.V. All rights reserved. december 2008, Vol. 66, No. 11 474 some researchers do not use it. Instead, they calculate the Calculations MAP from the SBP and DBP displayed by the device with The last three and five BP measurements, respectively, the formula DBP + 1/3 PP.8,9 Of note, some devices do not were used for analysis. SBP, DBP and MAP were retrieved report MAP. from the printed output lists. Calculated MAP was derived It is unknown if the measured MAP and calculated MAP from SBP and DBP using the formula DBP + 1/3 PP. PP are similar. In this study we compared the measured and was calculated by SBP – DBP. In each individual there calculated MAP obtained by two different oscillometric were three and five pairs of calculated and measured MAP, BP measurement devices in two study groups. Our data respectively. To obtain the average difference per subject, suggest that measured and calculated MAP cannot be used the calculated MAPs were subtracted from the measured interchangeably. MAPs and these values were averaged (MAPm-c ). For paired comparisons we used the Wilcoxon signed-rank test, for unpaired comparisons we used the Mann-Whitney test. Methods The MAPm-c was correlated with several variables using Spearman’s analysis. Multiple logistic regression was used For this study we used archival BP data obtained with an to determine factors independently related to MAPm-c. oscillometric BP measurement device in two different The analyses were done for the two groups separately. To patient cohorts. evaluate the possible effect of the type of device, we also Firstly we retrieved recordings of oscillometric BP analysed the overall dataset. measurements performed at our research unit in All data are presented as means (±SD) or medians (range) persons who were evaluated in the course of a screening when appropriate, All statistics were performed using programme for the detection of kidney disease. Participants SPSS software, version 14.0 (Chicago, IL). Differences were filled in a questionnaire on medication use. Body weight considered significant with p value <0.05. and height were measured, BMI was calculated. Blood pressure was measured using an automated oscillometric device (Welch Allyn 5300P) while subjects were in a sitting ResUlTs position with the arm supported at heart level. Five BP readings were done at five-minute intervals. Group 1 For the second analysis we used BP recordings of patients BP readings were available for 242 subjects. Their with kidney disease participating in a research programme clinical characteristics are shown in table 1. We observed on markers of progression of glomerular disease.10,11 a significant difference between measured MAP and In these patients approximately ten consecutive BP calculated MAP (p<0.001). The nonparametrical readings were performed at three-minute intervals with an distribution of the MAPm-c is shown in figure 1A. The automated device (Dinamap 487210, Critikon Tampa FL). median MAPm-c amounted to -1.8 mmHg (p<0.001), but In these patients BP was also measured by an experienced the difference can be large in individuals (range -5.7 to nurse using a sphygmomanometer. This ‘office’ reading 12.9 mmHg). The MAPm-c was slightly but significantly always followed the automated measurement. The use different in men and women (p=0.008). The median of an ACE inhibitor, β-blocker, diuretic agent or calcium MAPm-c in this group was -2.0 mmHg in male subjects antagonist was recorded. and -1.7 mmHg in female subjects. Correlations of MAPm-c Table 1. Baseline characteristics Group 1 (n=242) Group 2 (n=134) Variables All Male female All Male female (n=98) (n=144) (n=85) (n=49) Age (years) 50±13 55±12 47±12 50±14 51±14 50±15 BMI (kg/m2) 25.7±4.7 26.3±3.8 25.3±5.2 27.0±4.7 26.8±4.2 28.2±5.5 Systolic blood pressure (mmHg) 121.8±14.0 126.8±13.8 118.4±13.2 131.9±25.2 132.1±25.6 131.6±24.8 Diastolic blood pressure (mmHg) 74.9±9.7 78.3±9.3 72.6±9.4 79.2±12.2 79.3±12.9 79.2±10.9 Measured mean arterial pressure (mmHg) 90.0±10.5 93.6±10.7 87.6±9.7 98.1±15.9 99.2±16.8 96.2±14.3 Antihypertensive treatment* (%) 16.1 14.3 17.4 86.6 89.4 81.6 Values are expressed as means ± SD or %. *Antihypertensive treatment: use of calcium entry blockers/ACE inhibitor or angiotensin II receptor blocker/diuretic/β-blocker. Kiers, et al. Differences between measured and calculated MAP. december 2008, Vol. 66, No. 11 475 figure 1A. Histogram showing the difference between with age, BMI, SBP, DBP, measured MAP, pulse pressure measured and calculated mean arterial pressure in and heart rate are shown in table 2. Only DBP and SBP 242 subjects showed a significant, but weak correlation with MAPm-c. In multivariable analysis it appeared that sex was not 80 independently related to MAPm-c. Both SDP and DBP were 70 significantly related to MAPm-c. 60 50 Group 2 40 BP readings were available for 134 patients (table 1). In 30 this group we also observed a significant difference 20 between measured MAP and calculated MAP (p<0.001), Number of patients 10 with again a nonparametrical distribution of the MAPm-c 0 (figure 1B). The median MAPm-c amounted to +1.67 mmHg -5 0510 15 (p<0.001), with a large range between individuals (-15.3 MAP ∆ m-c to 28.2 mmHg). Figure 2 describes the relation between MAPm-c and the mean MAP. In this group there was no Blood pressures recorded with Welch Allyn 5300P oscillometric difference in MAPm-c between male and female subjects. device. Clearly, distribution of MAP is skewed. m-c All BP indices and heart rate correlated with MAPm-c, although none of these correlations were very strong (table 2). In multivariable analysis only age (p<0.001) was figure 1B. Histogram showing the difference between independently related to MAPm-c.